Assistance - Calling all Samsung EVO owners

Associate
OP
Joined
19 Oct 2002
Posts
1,460
Just for interest I get exactly the same heatbeat effect on my SSD in clean disc space.

I've sent some more information to Abarrass for him to pass on to Samsung with some resource mon shots while running the benchmark and acronis true image comparing the EVO to the Sandisk in my system showing that not only do each program show issues windows shows the issue as well but only on the EVO drive.
 
Last edited:
Associate
Joined
18 Oct 2002
Posts
1,765
Location
Kent, UK
I think it started with the latest firmware. I don't recall seeing these readings with previous firmwares

Nope. At the point I posted my info/benchmarks on the first page of this thread I was running an older firmware. I have since updated to the latest to see if it made a difference (it didn't... :( )
 
Associate
OP
Joined
19 Oct 2002
Posts
1,460
An update for everyone -

Samsung are going to send me a new drive to test in my machine to see if I get exactly the same result. At this stage I am not sure if that drive is another EVO or a PRO.

I have advised that I don't expect the PRO to show the issue as we have seen from others in this thread but lets see.

To be honest I'd prefer to get the EVO to see if I can replicate the problem to help resolve this though I won't refuse a minor upgrade. Both path's will show something.

I have advised them that regardless we need to fix this for everyone as many others of you have the issue not least my partner who has the same model of drive.

If this drive fixes the problem they want my old drive back which I am happy to secure ERASE and it will go to Samsung HQ for testing.

It makes me think they can't yet replicate what we are seeing which just seems mad!

My plan -

Get new drive. Acronis old drive to new drive. Test drive with HDTune and acronis over a couple of days.

This is what happened to the last EVO so lets try it and see what comes.

If it does fix it I can see they will get a lot of RMA's coming their way.
 
Last edited:
Associate
Joined
24 Aug 2003
Posts
317
If they want a drive that has not been secure erased and will ship me a replacement non EVO drive beforehand so i can acronis my existing one, then they can have my drive without a secure erase happening.

Barring a firmware fix coming out very soon, they are going to receive my drive as a RMA anyway.
 
Associate
Joined
20 Aug 2020
Posts
2,038
Location
South Wales
Here is my benchmark with an 8mb block size on the 16th August.

8mbBlockSize_zpse07df24b.jpg


Here is my 8mb block size benchmark just now, just 9 days later :eek:

8mb_zps669052b4.jpg


Nothing on my system has changed during the 9 days apart from using an extra 10gb of space.

I am using around 50% of the drive currently. This was a clone from a 2tb Seagate that was defragged before cloning. Overall though, the performance seems fine. There is no noticable diffence in use now than when I installed it.
 
Man of Honour
Joined
20 Sep 2006
Posts
34,043
I've just run an old M4 256GB through HD Tune and I am seeing similar.

Should I be looking to secure erase this drive and start over?

nTQn5J7.png

My M500 doesn't look great either.

FumowHw.png
 
Last edited:
Associate
Joined
2 Nov 2009
Posts
2,436
Location
Brum
I am due to carry out a "spring clean" (wipe everything and reinstall Windows).

If and when I do, I'll test again. I will do a secure erase before the reinstall.
 
Associate
OP
Joined
19 Oct 2002
Posts
1,460
Okay so big update -

I received my new drives identical to the old ones.

I installed one new drive on my PC and benchmarked it with a blank MBR NTFS format default allocation size

NEW_EVO_NO_DATA_zps15fddf95.jpg


So nice and tidy and performing as intended.

I then benchmarked the old EVO OS drive to show the issue.

OS_EVO_BEFORE_zpsbe11c056.jpg


As you can see very very unhealthy EVO.

So I cloned the OS drive directly to the new drive using Acronis True image home outside the OS. It took 30 mins approx. Which was slow and due to the poor source read performance of the unhealthy old EVO.

Then I benchmarked the newly cloned EVO os drive.

NEW_EVO_AFTER_CLONED_OS_zps22e560f6.jpg


Seems the cloning process halved the drives performance and introduces quite large heartbeat spikes. Not so good as performance has dropped about 200MB from the clean drive.

I then Manually trimmed the drive to see if this would help after all I just wrote 60GB+ to the drive.

NEW_EVO_AFTER_FIRST_MANUAL_TRIM_zpsd32dafb1.jpg


Nothing.. trim did absolutely nothing, fair enough.

Okay lets hammer the drive using the Samsung benchmark which writes a couple of gig every time you run it to encourage a drive TRIM and a bit more wear. To do this I benched it 5 times with rapid on. Disabled rapid and did it 5 more times. I then re-enabled rapid and ran a manual TRIM to try to clean up the drive.

After the 10 benchmarks runs without rapid mode

NEW_EVO_AFTER_10_BENCHMARK_RUNS_RAPID_DISABLED_zps46b4b4cb.jpg


Not much changed...

After the 10 benchmarks runs with rapid mode enabled and a manual TRIM

NEW_EVO_AFTER_10_BENCHMARK_RUNS_RAPID_RE_ENABLED_MANUAL_TRIM_zpsd52886ae.jpg


Yeah that made not much difference at all.


Thoughts so far - clearly cloning the drive across kills the performance off the bat and I just can't explain the heartbeat/stairs effect at all. I'm not seeing the sub 50MB/s of the bad drive yet though but this is day 0.

I'll keep you posted.

Worth a mention is the acronis drive copy time for the new EVO to my sandisk is back to where it should be at around 14mins from the 30+ it was reporting on the bad drive.

One final shot to prove it's not the PC that causes the heartbeat effect or the performance issues here is my sandisk ultra plus which is about the same full and 9 months old.

SANDISK.jpg
 
Last edited:
Associate
Joined
2 Nov 2009
Posts
2,436
Location
Brum
Interesting.

I'm pretty certain that my current OS install was restored from an image (using EaseUs Todo - with "Optimize for SSD" selected).

Earlier today I carried out a disk/partition backup, so I suppose I could do a clean install and see if the benchmarks look any better.
 
Back
Top Bottom